A Genetic Algorithm for the Estimation of Nonidealities in Continuous-Time Sigma–Delta Modulators

In this brief, a novel approach to using genetic algorithms (GAs) for estimating nonidealities in continuous-time sigma-delta (ΣΔ) modulators during runtime is presented. Since various nonidealities decrease the performance of ΣΔ modulators even up to instability of the circuit, there have been several publications for estimating these parameters in order to calibrate the analog-to-digital converter. Most of these techniques focused on individual nonidealities only. An unscented Kalman filter was previously presented as being able to estimate some of the most aggressive nonidealities in a joint fashion. However, the computations required in the Kalman filter are highly challenging. Hence, a heuristic algorithm, which is generally very simple from the mathematical point of view, is proposed as a better choice for the previously mentioned application. It will be shown that a specific form of the GA is able to estimate several nonidealities concurrently, thus allowing for system identification for later calibration of the modulator.